Testing the scale dependence of plant community assembly processes using imaging spectroscopy

The scale of observations can have large effects on inferences about community assembly processes, but researchers rarely have the resources to sample many different scales. Imaging spectroscopy, a passive optical remote sensing technology, enables data acquisition across many scales, opening the door to replicating analyses across multiple spatial grains and extents. We applied hyperspectral data acquired by imaging spectroscopy to test the effects of scale on inferences about tree community assembly processes along an elevational gradient in Québec, Canada. We found strong scale dependence in how much variance in spectral composition was attributed to environmental versus spatial predictors. With increasing sampling grain (plot size), environmental selection was increasingly important relative to ecological drift. By contrast, the effect of environmental extent (elevational range) was idiosyncratic. Limiting analyses to only high-elevation plots had no influence on the magnitude of variation explained (compared with the entire gradient), while considerably less variation was explained in low-elevation plots, a result consistent with theory comparing harsh and benign environments. Our study demonstrates the utility of imaging spectroscopy to permit the study of a wider range of scales than previously possible, and so to help achieve a more scale-integrative community ecology.